Is ChatGPT becoming a lazy AI Assistant?

Are AI Agents and AI Assistants becoming useless

I think if the Agents stop helping and ask us to do everything on their behalf they become a bit useless, but I prefer the term watered down it is more obvious.

I guess that since I am not a corporation I can not have the Enterprise Version and since I do not have anyone else to share the service with I can not get the Team plan.

The eternal frustration of using ChatGPT

I do not know how a corporation would manage to use the Chat Assistant if they where to have such results all the time… I am a paid user because I want to have the full benefits of the AI Agent out of the box, not to use complicate prompting techniques to make the AI do what I know he is capable of doing…

I would like to know if it is possible to add a feature where ChatGPT can browse the internet because I always need to do all the job myself and I am unsure why it is so complicated:

First Attempt

Screenshot_20240122_123708

Second Attempt

Screenshot_20240122_123714

Third Attempt

So I decided to read the documentation myself and I extracted the information I wanted the Chat Assistant to help me understand I asked him this:

[…] then the callback API is the one which have the most items if I made no mistake it will be around 48 please ingest them and then produce the comparative list add a last column going from 1 to the number of rows at the last position please and list each of the 3 APIs (paradigms) pairing the elements that are the same across multiples of them LOOK:
[the third list of elements]

Then I got that

Screenshot_20240122_124233

I then stated to be a bit less polite insulting the Agent by comparing him as a human…

what the heck this is not what I want why are you lazy you are not a real person remember that you are an assistant
what the heck this is not what I want why are you lazy you are not a real person remember that you are an assistant

So I got to the 20 items limit

Screenshot_20240122_124540

Because I am not the kind of person who believes in conspiracy theories I would not say that it is done on purpose to make use more questions to get a same result. I do understand that it might be to limit the costs (with shorter outputs) rather that increasing the spending (by forcing us to use more of our questions per period of time)…

In fact if anything me asking 3 to 4 time to get something is going to put a toll on the infrastructure of the LLM I would like to know how to prompt my questions in a way that the AI would not be Lazy all the time…

The full conversation can be found bellow

I have not used special instruction in that specific conversation and I would like to have the help of the community to mitigate this issue until it if fixed hopefully by OpenAI if it is a thing that can be fixed.

I have not used the AI Assistant to help make correction to my text because it would have take me too long to write my text… (Yes I know it is that bad indeed)